Re: Contributing to Hurd

As usual, I forgot the attachments :)

Missatge de Joan Lledó <[email protected]> del dia dg., 3 de
març 2019 a les 17:25:
>
> Hi,
>
> > Joan, where do think something could be missing?
>
> Maybe a good start is to read and understand all the code of the
> translator and explain what every function does. That could be useful
> for others who want to contribute to the Hurd, not only on the lwip
> translator. Back in my GSoC I took some notes about pfinet, lwip,
> glibc and ip6. Attached are the notes but please Sooraj, take into
> account these are personal notes, not intended to be published and the
> statements I make in these documents *may be wrong*. I hope these
> notes help you to complete your documentation.
>
> Another thing you could do as part of your thesis is to make
> performance tests on both lwip and pfinet translators, to check which
> is faster and what's the difference. I never tested this, but I
> remember Justus did some quick tests and said the lwip translator was
> far slower than pfinet. It would be nice to confirm it and figure out
> why.
>
> >Also, are there perhaps some features that Sooraj could work on?
>
> If I remember correctly, poll() and ppoll() on glibc end up calling
> the io_select operation. Maybe a good exercise would be to add a new
> 'poll' operation to the pfinet interface, b/c it's not a big change
> but it requires changes on glibc, the MIG interface and the
> translator, so it's a nice exercise to have the full picture of how
> the Hurd works.
>
> What do you think?